The research on banknote authenticity discrimination analysis algorithm based on wavelet transform features
Sijian Zhao · Journal of Artificial Intelligence Practice · 2024
In the authenticity identification of banknotes, features such as variance, skewness, kurtosis, and entropy of the images transformed by wavelet are used. This paper combines distance discriminant analysis, Fisher discriminant analysis, and Bayesian discriminant analysis for discrimination analysis. Variance can measure the texture complexity and grayscale level variation in the image, skewness is used to evaluate the symmetry and deviation from the normal distribution of the image, and kurtosis can measure the texture structure and grayscale level concentration. The entropy of the image reflects the complexity and uncertainty of the image. These features can be used to distinguish genuine banknotes from counterfeit ones.